This project addresses a key methodological challenge in bilingual child language research: the automatic detection and transcription of mixed-language speech (e.g., code-switching) from daylong LENA recordings. Current manual transcription is labor-intensive and time-consuming; AI tools offer promising solutions. We will conduct a literature review and feasibility study to determine if AI transcription tools can accurately detect and transcribe code-switching in naturalistic bilingual child-parent interactions.
Many bilinguals combine or mix languages when they speak. We aim to find out if children have difficulties learning language from mixed sources, how and why children mix, and if language mixing is different for children with and without a Developmental Language Disorder. You can read more about this research project via https://www.uu.nl/en/research/calm.